Keywords
PTMC - preoperative thyroglobulin - radioiodine therapy - prognosis
Introduction
Papillary thyroid microcarcinoma (PTMC) is commonly referred to as papillary thyroid
carcinoma (PTC) with a maximal diameter of≤10 mm. Despite its prevalence, most PTMCs
behave in an indolent way and typically have good prognoses [1]. For those low-risk PTMC, active
surveillance (AS) or other non-surgery methods were accepted as an alternative
management option according to the 2015 American Thyroid Association (ATA)
guidelines [2]. For PTMC patients with lymph
node metastasis (LNM), or extrathyroidal extension (ETE), traditional treatment
including surgery and postoperative radioactive iodine therapy (RAIT) will be
implemented, if necessary [2]. However, poor
response to RAIT and risk of recurrence still exists in PTMC after routine therapy
[3]
[4]
[5]. Identifying predictors in
risk stratification of PTMC with poor prognosis will provide great supports for the
decision making of AS or treatment management for PTMC patients. Up to date, only
few risk factors like age<40, male gender, multifocality, and calcification have
been reported to be associated with high-risk clinical features [6] and therefore more efforts are needed to
uncover some other risk factors for early identifying PTMC with poor prognosis.
Thyroglobulin (Tg) is synthesized in thyroid follicular cells under the regulation of
thyroid stimulating hormone (TSH) [7]. In
healthy individuals, serum Tg is barely detectable. The level of Tg was applied in
an attempt to the field of thyroid cancer diagnosis at first but it failed due to
the overlap between benign and malignant thyroid tissues [2]. However, recent findings showed an
association of pre-Tg with the poor prognosis of thyroid cancer like LNM, distant
metastasis, and so on [8]
[9]
[10].
Besides, our recent study provoked a prognostic value of pre-Tg in classifying
patients with risk of radioiodine refractory (RAIR) [11]. Interestingly, we also noticed a relative high level of pre-Tg in a
small number of PTMC patients with RAIR [11].
It is well-known that Tg level may be influenced by certain factors like antibodies
of Tg (TgAb) and thyroiditis like Hashimoto’s thyroiditis (HT) [2]. Whether pre-Tg has a predictive value in
PTMC after the exclusion of these interference factors remains unknown.
Here, we compared the pre-Tg level in patients with or without radioiodine therapy
(RAIT), or with single and multiple cycles of RAIT. Furthermore, the association of
pre-Tg level and the response to RAIT in PTMC were investigated. Our study provoked
pre-Tg as a novel predictor for PTMC prognosis and may shed light on the choice of
personalized management of PTMC in the future.
Subjects and Methods
Study participants
This retrospective study was conducted in Jiangyuan Hospital affiliated to
Jiangsu Institute of Nuclear Medicine. Preoperative thyroid ultrasonography
(US), and thyroid function test were done routinely. Based on the clinical
decision-making framework suggested by Brito et al., patients were evaluated for
the options of management [12]. For
low-risk PTMC, both lobectomy and active surveillance were suggested in our
hospital.
Exclusion criteria
There were 4482 PTMC patients who underwent thyroid surgery from January 2015 to
December 2019 in our hospital. After the exclusion of patients with
anti-thyroglobulin antibody (TgAb) positive, thyroiditis, history of fine needle
aspiration (FNA), and less than 1-year follow-ups, etc., 788 PTMCs with full
records including pre-Tg, TgAb, thyroid stimulating hormone (TSH), and
post-surgery Tg were enrolled. The follow-up time course was decided as the
period from surgery to the index date. Each index date for each patient was
referred to the latest laboratory results from January 2015 to December 2022.
The median follow-up was 67.55 (range 49.53–80.64) months.
Surgery and RAIT
Lobectomy was conducted in 496 PTMCs, and total thyroidectomy was done in 292
patients. A total of 107 patients were treated with RAI after undergoing total
thyroidectomy. ATA risk stratification was determined based on 2015 ATA
guidelines [2]. For partial PTMC patients
with intermediate risk and high risk, an empirical dose of 100–250 millicurie
(mCi, 3.7–9.25 gigabecquerel, GBq) was administrated. All patients were
recommended to have 6- and 12-month follow-ups after RAIT.
The response to RAIT was evaluated based on 2015 ATA guidelines [2]. Briefly, the response to RAIT was
determined based on their imaging and serologic results. Then it was subscribed
into excellent response (ER), and non-excellent response (NER), which includes
indeterminate response (IDR), biochemistry recurrence (BIR), and structurally
recurrence (SIR). After 1-year follow-up [13.2 (10.0–15.1) months], 76/107
patients exhibited an ER, 15/107 patients developed a IDR, and 16/107 patients
developed a BIR. At the last follow-up [51.5 (32.3–66.9) months], 82/107
patients showed ER, 16/107 patients showed IDR, 8 patients developed BIR, and
1/107 patient developed SIR. The only SIR patient was classified in the BIR
group in the current study.
Variables
Electronic clinical documentation, pathologic information, laboratory data,
surgical reports, radiology reports, and molecular testing results were obtained
from Jiangyuan Hospital electronic information system if possible. With the
approval of the institutional ethics committee (YL202207), the clinical data of
patients were carefully reviewed.
Age was recorded on the date of diagnosis, and the cut-off age for risk
stratification was set at 55 years according to the Eighth Edition of the
American Joint Committee on Cancer [13].
All involved PTMC were classical PTMC. Gross ETE was defined as the macroscopic
extension of the primary tumor into the perithyroidal tissues including strap
muscles, trachea, recurrent laryngeal nerve, and blood vessel. Positive LNM was
defined as the number over 5 LNMs [2].
Tg, TgAb, TPOAb, and TSH of patients before surgery were measured by a clinical
laboratory using Cobas 8000 modular analyzer series with indicated
immunochemiluminescence assay kits according to the manufacturer’s instructions
(Roche, Mannheim, Germany). The quantification limits of Tg and TgAb, TPOAb
measurements were from 0.04–500 ng/ml, 0–115 IU/ml and 0–34 IU/ml, respectively.
The normal or reference range for TSH was 0.35 to 5.5 mIU/l in this study. The
normal range for TgAb was below 115 IU/ml, and the normal range for TPOAb was
below 34 IU/ml.
Statistical analysis
Normally distributed continuous variables are expressed as mean±standard
deviation (SD) and compared using the independent-sample t-test.
Non-normally distributed continuous variables are presented as median with
interquartile range (IQR) and compared using the Mann–Whitney U-test.
Categorical variables were presented as numbers with percentages and compared
using Chi-square and Fisher’s exact methods. For receiver operating
characteristic (ROC) analysis, GraphPad Prism 8.2.1 was used to draw a ROC curve
and the optimal cut-off value of pre-Tg for predicting the prevalence of RAIT
was selected. For multivariate analysis, binary logistic regression with the
forward LR method was performed for all demographic and pathologic variables. A
two-sided p<0.05 was considered statistically significant.
Results
High level of pre-operative Tg was found in PTMC with repeated RAIT
This study included 788 PTMC patients (590 women, and 198 men) according to the
inclusion criteria ([Fig. 1]). A total of
91 patients received single RAIT and 16 patients received multiple cycles of
RAIT. The patients’ baseline characteristics are shown in [Table 1]. All clinical pathologic factors
showed significant differences between groups of patients with or without RAIT
(p<0.001). When comparing patients with single or multiple cycle of RAIT, no
clinical factors showed difference except factors of LNM (p=0.03) and LNM number
over 5 (p<0.001).
Fig. 1 Flow chart of the study. The cohort included 4482 patients
with PTMC. After the inclusion/exclusion criteria were applied, 788 PTMC
patients were finally involved.
Table 1 Clinicopathological characteristics of the
papillary thyroid microcarcinoma patients.
Variables
|
No RAIT
|
Single RAIT
|
Repeated RAIT
|
p-Valuea
|
p-Valueb
|
Number of cases
|
681
|
91
|
16
|
|
|
Median age (years)
|
44 (35, 52)
|
46 (38, 52.5)
|
37 (32.25, 44)
|
|
|
Age at diagnosis (years)
|
|
|
|
0.922
|
0.22
|
<55
|
569 (83.6)
|
74 (81.3)
|
15 (93.8)
|
|
|
≥55
|
112 (16.4)
|
17 (18.7)
|
1 (6.3)
|
|
|
Sex
|
|
|
|
0.612
|
0.69
|
Female
|
512 (75.2)
|
67 (73.6)
|
11 (68.8)
|
|
|
Male
|
169 (24.8)
|
24 (26.4)
|
5 (31.2)
|
|
|
Maximum tumor size (cm)
|
|
|
|
<0.001
|
0.91
|
<0.5
|
254 (37.3)
|
16 (17.6)
|
3 (18.8)
|
|
|
≥0.5
|
427 (62.7)
|
75 (82.4)
|
13 (81.2)
|
|
|
Multifocality
|
|
|
|
<0.001
|
0.13
|
Yes
|
154 (22.6)
|
63 (69.2)
|
8 (50.0)
|
|
|
No
|
527 (77.4)
|
28 (30.8)
|
8 (50.0)
|
|
|
Gross ETE
|
|
|
|
<0.001
|
0.47
|
Yes
|
22 (3.2)
|
11 (12.1)
|
3 (18.8)
|
|
|
No
|
659 (96.8)
|
80 (87.9)
|
13 (81.2)
|
|
|
LNM
|
|
|
|
<0.001
|
0.03
|
Yes
|
215 (31.6)
|
70 (76.9)
|
16 (100.0)
|
|
|
No
|
466 (68.4)
|
21 (23.1)
|
0 (0)
|
|
|
LNM (n>5)
|
|
|
|
<0.001
|
<0.001
|
Yes
|
24 (3.5)
|
27 (29.7)
|
8 (50.0)
|
|
|
No
|
657 (96.5)
|
64 (70.3)
|
8 (50.0)
|
|
|
Tumor laterality
|
|
|
|
<0.001
|
0.34
|
Unilateral
|
479 (70.3)
|
13 (14.3)
|
13 (81.2)
|
|
|
Bilateral
|
202 (29.7)
|
78 (85.7)
|
3 (18.8)
|
|
|
BRAF V600E
c
|
|
|
|
0.418
|
0.18
|
Yes
|
77 (96.2)
|
7 (100)
|
3 (75.0)
|
|
|
No
|
3 (3.8)
|
0 (0)
|
1 (25.0)
|
|
|
ATA risk after surgery
|
|
|
|
<0.001
|
1.00
|
Low
|
399 (58.6)
|
0 (0)
|
0 (0)
|
|
|
Intermediate+High
|
282 (41.4)
|
91 (100)
|
16 (100.0)
|
|
|
Accumulated
131
I doses
|
0
|
120 (100–135)
|
300 (230–320)
|
–
|
<0.001
|
Median follow-up time (months)
|
64.3 (48.8–79.2)
|
78.7 (60.2–87.8)
|
73.9 (53.9–86.5)
|
|
|
a p-Value was compared between patients without RAIT and with
RAIT. Pearson Chi-Square value. b p-Value was compared
between patients with single and repeated RAIT. Pearson Chi-Square
value. c
BRAF mutation, partial data available,
n=91.
However, the level of pre-Tg was higher in the RAIT group. The median levels of
pre-Tg for each group were as follows: 10.63 (range 5.4–20.1) ng/ml in the
no-RAIT group, 16.5 (range 6.6–40.5) ng/ml in the RAIT group (p=0.0018).
Further, the level of pre-Tg was much higher in patients with repeated RAIT. The
median levels of pre-Tg for each group were as follows: 12.86 (5.19–27.8) ng/ml
in single RAIT group, and 40.67 (21.77–83.46) ng/ml in repeated RAIT group
([Fig. 2]).
Fig. 2 High level of pre-Tg in PTMC with repeated RAIT. a:
Scatter plots with histograms of pre-Tg levels in groups of patients
with or without RAIT. Mann–Whitney U-test, p=0.0018. b:
Scatter plots with histograms of pre-Tg in patients with 1 cycle or>1
cycle of RAIT. Mann–Whitney U-test, p<0.0001.
Higher pre-operative Tg was found in PTMCs with poor prognosis after
RAIT
Among 107 PTMC patients with RAIT, 82 patients showed ER, 16 patients showed IDR,
and 9 patients showed BIR at the last follow-up. The median value of pre-Tg was
11.8 (range 4.7–25.4) ng/ml in the ER group, 21.5 (range 9.5–75.0) ng/ml in the
IDR group and 59.5 (range 39.4–143.4) ng/ml in the BIR group (p=0.0003) ([Fig. 3a]).
Fig. 3 Higher level of pre-Tg in patients with IDR and BIR.
a: Scatter plots of pre-Tg in patients with excellent
response (ER), indeterminate response (IDR) or biochemical incomplete
response (BIR). Mann–Whitney U-test, p=0.0003. b: ROC
curve for pre-Tg in PTMC patients with NER, p=0.0006.
To evaluate the predictive performance of pre-Tg for RAIT outcome (ER and NER) in
PTMCs, the ROC curve was fitted and the corresponding area under curve (AUC) was
0.73 (0.61–0.85, p=0.0006). Setting the cut-off point of pre-Tg at 16.79 ng/ml
gained a maximum 80.0% (59.3–93.2%) in sensitivity and 61.0% (49.6–71.6%) in
specificity.
Pre-operative Tg was the independent predictor for poor prognosis of RAIT in
PTMCs
In univariate analysis, factors found with statistical significance in predicting
NER were as follows: LNM number over 5 (OR=2.96; 95% CI, 1.17–7.44; p=0.022),
bilaterality (OR=0.33; 95% CI, 0.12–0.94; p=0.039), and pre-Tg over 16.79 ng/ml
(OR=6.25; 95% CI, 2.13–18.33; p<0.001). When taken these factors in further
multivariate analysis, only pre-Tg over 16.79 ng/ml (OR=5.91; 95% CI,
1.96–17.83; p=0.002) was the independent predictor for NER in PTMC patients with
RAIT ([Table 2]).
Table 2 Univariate and multivariate analysis of the risk
factors of NER in patients with RAIT.
Variables
|
Univariate analysis
|
Multivariate analysis
|
OR (95% CI)
|
p value
|
OR (95% CI)
|
p-Value
|
Sex (Male)
|
2.21 (0.85–5.72)
|
0.102
|
–
|
–
|
Age (≥55)
|
0.36 (0.08–1.68)
|
0.193
|
–
|
–
|
Tumor size (≥0.5 cm)
|
1.18 (0.35–3.93)
|
0.793
|
–
|
–
|
Multifocality (Yes)
|
0.45 (0.18–1.12)
|
0.087
|
–
|
–
|
Gross ETE (Yes)
|
2.02 (0.61–6.73)
|
0.248
|
–
|
–
|
LNM (n>5)
|
2.96 (1.17–7.44)
|
0.022
|
2.31 (0.83–6.41)
|
0.109
|
Bilaterality (Yes)
|
0.33 (0.12–0.94)
|
0.039
|
0.43 (0.13–1.41)
|
0.164
|
Pre-Tg (≥16.79 ng/ml)
|
6.25 (2.13–18.33)
|
<0.001
|
5.91 (1.96–17.83)
|
0.002
|
OR: Odds Ratio; pre-Tg: preoperative Tg.
Discussion
In the current study, we showed that a high level of pre-Tg was found in PTMC
patients with RAIT, repeated RAIT and poor response to RAIT. In comparison of PTMC
patients with ER at the last follow-up, pre-Tg level was higher in patients with
NER. All these results provided pre-Tg as a feasible biomarker in evaluating PTMC
patients with the risk of poor prognosis. Consistently, more and more studies showed
that pre-Tg was closely related with the prognosis of thyroid cancer. Our previous
study showed that high level of pre-Tg>70.05 ng/ml in PTC was a prognostic marker
for the development of RAIR [11]. Likewise,
high serum pre-Tg level has been provoked to be associated with tumor metastasis in
PTC [9]
[14]
[15]
[16]. In FTCs, pre-Tg was also regarded as a
valuable predictor for distant metastasis [15]. Although pre-Tg can not better discriminate benign nodule from thyroid
cancer, it is more acceptable that the level of pre-Tg was a promising biomarker for
thyroid cancer prognosis.
Risk factors like age<40, male gender, multifocality, and calcification were found
to be correlated with LNM in PTMC [6].
However, the occurrence of LNM was not closely related with the prognosis and most
small metastatic cervical LNs remained stable during AS in PTMC [17]. Similarly, our results showed that LNM
number over 5 was not an independent risk factor for PTMC prognosis but we found it
was correlated with the response to RAIT. In the current study, our conclusion might
be influenced by the small sample size and more efforts were needed to verify the
correlation of LNM with PTMC progression.
Whether patients can benefit from repeated RAIT is still in controversy. One report
pointed out that a second RAIT hardly benefited thyroid cancer patients [18]. Similarly, we also found a high
consistency in the response to RAIT at the time of one-year follow-up and the last
follow-up. Limited number of patients were benefited from the second or more cycles
of therapy. This observation may be influenced by the time of Tg measurement and
needed to be further confirmed.
In recent years, active surveillance (AS) has been endorsed as an alternative
management approach for upfront surgical treatment in low-risk PTMC [19]. It markedly decreased the side effects
like alterations in voice associated with damage to recurrent and superior laryngeal
nerves, hypoparathyroidism caused by the surgery [20]. Thus, the 2016 Chinese medical expert consensus recommended AS for
patients with low-risk PTMC [21]. However,
like in our cohort, there were still a considerable number of PTMC patients with
lobectomy even they have a good prognosis within the 5-year follow-up. Barriers were
caused by the complexity in the decision-making process and the nature fear of
cancer in patients [22]
[23]
[24].
Enhancing the consensus among clinicians and informational supports for patients
will largely improve the ratio of PTMC with AS when applicable. Meanwhile, a
feasible biomarker to distinguish the aggressive PTMC will strengthen the confidence
of clinicians and patients when facing the choice making from AS and upfront
surgery.
BRAF V600E mutation molecular test has been widely used in identifying malignant
thyroid cancer [2]. It may be promising to
combine the use of BRAF V600E and pre-Tg in predicting PTMC patient prognosis. By
distinguishing PTMC patients with poor prognosis, it may also help in the decision
making of AS for patients with fear of disease progression.
Limitations
Our study has several limitations. First of all, due to the low prevalence of
high-risk features in PTMC, a limited eligible sample size of PTMC with repeated
RAIT and poor response to RAIT was enrolled in our cohort.
Second, the cycles of RAIT received by patients and the response to RAIT used in
the present study cannot best reflect the recurrence and progression of PTMC. It
still needs more studies concerning the real state of recurrence and metastasis
in PTMC, though its incidence is low, to verify the prognostic role of
pre-Tg.
More importantly, the underlying mechanism that why high pre-Tg level correlated
with poor prognosis of PTMC remained elusive. We have excluded the interference
of thyroiditis, TgAb, and other possible factors. However, we only found a weak
correlation of pre-Tg and LNM number over 5 (data not shown). One study has also
pointed out the correlation of pre-Tg and LNM by establishing a nomogram model
for the prediction of cervical and lateral LNM based on serum pre-Tg levels
[25]. It still needs more efforts to
uncover the mechanism leading to the upregulation of pre-Tg in patients with
risk of poor prognosis.
Also, the selection bias could not be neglected, and more studies were needed to
confirm our conclusion.
In summary, our results showed a high level of pre-Tg in patients with poor
prognosis in PTMC. It can offer a new prognostic marker for PTMC with RAIT.